Host Computer (Z840) Recon GPU Card Replacement

Prerequisites

Table 5. Safety

POTENTIAL FOR INJURY TO SERVICE PERSONNEL
THE HOST COMPUTER CAN WEIGH IN EXCESS OF 55 POUNDS.
DEPENDING ON YOUR HEIGHT AND ACCESS TO EQUIPMENT REMOVAL/INSTALLATION, LIFTING ASSISTANCE MAY BE REQUIRED. USE YOUR JUDGMENT TO DETERMINE IF LIFT ASSIST OR A SECOND PERSON MAY BE REQUIRED TO ASSIST IN REMOVAL AND INSTALLATION.

TO PREVENT DAMAGE TO COMPONENTS WITHIN THE Z840 COMPUTER, OBSERVE THE FOLLOWING ESD PRECAUTIONS:
WORK ON A STATIC-FREE MAT.
WEAR A STATIC STRAP TO ENSURE THAT ANY ACCUMULATED ELECTROSTATIC CHARGE IS DISCHARGED FROM YOUR BODY TO GROUND.
CREATE A COMMON GROUND FOR THE EQUIPMENT YOU ARE WORKING ON BY CONNECTING THE STATIC-FREE MAT, STATIC STRAP AND PERIPHERAL UNITS TO THAT PIECE OF EQUIPMENT.
FOR FURTHER INFORMATION REGARDING ESD PROTECTION, REFER TO THE SAFETY CHAPTER IN THIS PUBLICATION

Overview

This procedure shall be followed when replacing the Recon GPU card in the Host Computer (Z840).

Shutdown system. Select one of the following methods to Power OFF the Console:
- If Applications are up, click on the Shut Down button on desktop display and select Shutdown.
- If Applications are down, open a Terminal Window. Type: halt , then press ENTER.
- When halt command has finished, power Off the console at the front panel switch.
Apply LOTO. See Equipment Service - Lockout-Tagout-PPE procedure.
Remove Front and Top covers. Refer to the following procedure.

NIO64: Refer to Replacement → Console (NIO64) → RIO / NIO64 Console Cover Removal and Installation

GOC6.6: Refer to Replacement → Console (GOC6.6) → Console Cover Removal and Installation
Remove the host computer from console chassis. Refer to the following procedure.

NIO64: Refer to Replacement → Console (NIO64 Z840) → NIO64 Host Computer (Z840) Replacement

GOC6.6: Refer to Replacement → Console (GOC6.6) → GOC6.6 VCT Host Computer (Z820/Z840) Replacement

Open the host computer side access panel.
Remove the Expansion Card Support.

Figure 2. Z840 Airflow Guide and Expansion Card Support
Disconnect the power cable from the Recon GPU card.
Replace the existing Recon GPU Card with new one.

note:
Lift up the card latch when removing the card. (See Figure 4)

Figure 3. Z840 Component Location

Figure 4. PCI card Latch
Connect the power cable to the Recon GPU card.
Install the Expansion Card Support and close the Side Access Panel.

Install Host computer into Console chassis. Refer to the following procedure.

NIO64: Refer to Replacement → Console (NIO64 Z840) → NIO64 Host Computer (Z840) Replacement

GOC6.6: Refer to Replacement → Console (GOC6.6) → GOC6.6 VCT Host Computer (Z820) Replacement
Reconnect all cables removed earlier to the Z820/Z840 computer.
Install Console covers. Refer to the following procedure.

NIO64: Refer to Replacement → Console (NIO64) → RIO/NIO64 Console Cover Removal and Installation

GOC6.6: Refer to Replacement → Console (GOC6.6) → Console Cover Removal and Installation
Remove LOTO on console.

Perform the Functional Checks → System Scanning Test instructions from the procedure list.
Check the GPU card installed. Open a shell, then type:
{ctuser@hostname} ls /proc/driver/nvidia/gpus | wc —l
- If “2” displays, the Recon GPU card is installed.
- If “1” displays, the Recon GPU card is un-installed.
Ensure GPU ECC state is ON:
1. Open a Uni shell and log on as root.
2. Type: su - [ENTER].
3. Type the root password [ENTER].
4. Type: nvidia-smi [ENTER]
5. Check GPU ECC status as below:
  - If the GPU ECC is ON, below is what the output would look like (boxed in green):
    Figure 5. GPU ECC - ON
  - If the GPU ECC is OFF, below is what the output would look like (boxed in red):
    Figure 6. GPU ECC - OFF
6. How to turn ECC back on
  - •Type: nvidia-smi -g 0 --ecc-config=1 [ENTER]
  - •ill show that ECC is enabled and a reboot is required:
    Figure 7. Reboot Message
  - After reboot, check that the ECC is ON according to previous steps.
Perform the Functional Checks → System Scanning Test instructions from the procedure list.